Development of solar radiation database and its integration into solar process applications in Lesotho

Abstract

Solar energy is a viable alternative source of energy for socio-development of a developing country like Lesotho. Investment in solar process applications, requires a accurate solar radiation data for the successful implementation of solar process projects. However, in Lesotho measured solar radiation datasets are not sufficient both temporally and spatially as there are only seven solar radiation measuring sites, two of which are not reliably operational. This study solves the problem of the scarcity of solar radiation data in Lesotho, by developing a solar radiation database for the country. It has a primary objective of developing an accurate solar radiation database for Lesotho. This is achieved by merging ground measured solar radiation data with satellite – derived solar radiation data. The merged data is complimented by solar radiation data derived from sunshine duration data. Merging solar radiation datasets is important because ground measured data are sparsely distributed and cannot be interpolated accurately to represent solar radiation at any location. Although satellite – derived datasets are spatially continuous, they are not accurate as they are inferred from extra- terrestrial solar radiation modified with atmospheric models. As a result, each of the databases cannot be relied up unilaterally. Measured ground data is from five stations and sunshine duration derived solar radiation is form twelve stations. The improved database is validated using a leave one out cross validation technique. Its reliability in estimating ground solar radiation is tested by relative bias error (rBE), relative mean bias error (rMBE) and normalized root mean square error (NRMSE). Results show that the database is credible as it has a maximum error of 2.7 % which is comparable to other studies of similar nature in Africa. An interpolation tool, increases the accuracy and reliability of interpolated solar radiation as compared to manual reading of data from solar radiation maps. It has an accuracy of 99.54%. The improved database and interpolation tool can confidently be used in any solar application process design and sizing in the country.